一种改进的联合概率数据关联算法  

An Improved Joint Probability Data Association Algorithm

作　　者：骆荣剑[1] 魏祥[1] 李颖[1] LUO Rongjian;WEI Xiang;LI Ying(Chongqing Communication Institute,Chongqing 400035,China)

机构地区：[1]重庆通信学院,重庆400035

出　　处：《重庆理工大学学报（自然科学）》2018年第6期160-168,共9页Journal of Chongqing University of Technology：Natural Science

基　　金：国家自然科学基金资助项目(61272043)

摘　　要：针对联合概率数据关联算法在跟踪多机动目标时跟踪精度不高、计算量较大等问题,提出了一种新的联合概率数据关联算法。引入"当前"统计模型,并针对"当前"统计模型中机动频率和加速度方差不能自适应调整的问题进行了改进,实现了机动频率和加速度方差自适应。针对联合概率数据关联算法在跟踪多机动目标时,随着目标数的增多,算法计算量急剧增大的问题,提出了改进算法。改进算法避开了联合概率数据关联算法中由确认矩阵计算关联矩阵的过程,直接从确认矩阵计算关联概率。仿真实验结果表明:所提算法有效提高了多目标的跟踪精度,降低了算法计算量。Aiming at the problem that the joint probability data association algorithm is tracking the multi-maneuvering target,the tracking accuracy is not high and the computation amount is large. A new joint probability data association algorithm is proposed. This method first introduces the＂current＂statistical model and improves the problem that the maneuver frequency and acceleration variance cannot be adjusted adaptively in the ＂current ＂ statistical model,To achieve the motor frequency and acceleration variance adaptive adjustment. Then,when the multi-maneuvering target is tracked by the joint probability data association algorithm,with the increase of the number of targets,the computational complexity of the algorithm increases sharply. The improved algorithm avoids the process of calculating the correlation matrix from the validation matrix in the joint probability data association algorithm,and calculates the association probability directly from the confirmation matrix. The simulation results show that the algorithm proposed in this paper caneffectively improve the tracking accuracy of multi-targets and reduce the computational complexity.

关 键 词：联合概率数据关联 “当前”统计模型 多机动目标跟踪 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]